This proposal finalizes the design of a portable ankle robot as a therapeutic device to reverse foot drop and restore safer independent walking after stroke, then conducts a rigorous randomized clinical trial to establish clinical proof for this new market application. Currently, there are no therapeutic solutions for foot-drop, only assistive devices (braces, canes, electrical stimulation) that don't reverse the underlying ankle deficits. In contrast to arm robotics, lower extremity robotics have been largely unsuccessful in stroke patients to date, we posit because they do not emphasize volitional engagment for mediating motor learning. We pioneered modular ankle robotics technology, enabling human-robotics cooperative learning utilizing impedance control for deficit severity adjusted assist-as-needed in conjunction with adaptive control for precise timing of assistance to sub-events across the gait cycle. Our five positive clinical studies using this tethered research device show hemiparetic ankle motor learning that translates into improved over-ground gait. Most importantly, our randomized study shows that 18 sessions of ankle robotics integrated treadmill training (but not seated robot training) durably reverses foot drop, restores paretic leg propulsion, and promotes safer heel-first foot landing during unassisted walking, enabling 85% to self-discard their ankle braces or switch to a less dependent assistive device. Inspired by our findings that ankle robot mediated functional motor learning is most effectively conducted in the context of locomotor task-oriented training, we pursue finalization of a lighter weight and ergonomic portable ankle robot (AMBLE) that meets portability and clinical usability standards for untethered wear into regular mobility focused physical therapy (PT). Phase I produces a commercial portable ankle robot with desired portability (device mass, battery life); hardware design (FDA freeze)-microcontroller software (FDA compliant); and adequate clinical usability (client-therapist performance report card, auto-adjustable assistance) that meets key safety, stability, and comfort metrics (don-disengage time, slippage, part failure, skin abrasions/irritations) during over-ground walking and mobility tasks for full integration into physical therapy. Commercial units meeting FDA requirements will reach design freeze the first year (Phase 1); feasibility of which is supported by our precedent battery operated version already showing proof-of-function in stroke patients. Using AMBLE as a therapeutic device, optimized by the robot's intrinsic measurement capacities, we shift the robotics rehabilitation paradigm beyond repetitive robot guided exercise task practice, toward immersive PT integrated robotics (PTR). Phase II (3-yrs) is a blinded, randomized clinical trial investigating the hypothesis that 9 weeks (18 sessions) of PTR is more effective than PT alone to reverse foot drop as assessed by gait biomechanics (ankle angle at initial contact, peak swing ankle angle, number of heel-first strikes - % total steps, gait velocity; secondary outcomes- 3D gait biomechanics, daily ambulatory activity ) and blinded clinician assessment (dorsiflexion active range of motion, ankle muscle strength, assistive device needs). Our results will establish safety and initial efficacy for reversal of foot drop and restoration of biomechanically safer walking as a new market application filling an unmet need, and shift care toward PT integrated robotic exoskeletons across other impaired joints.